Log P is an important physicochemical characteristic and it can express the lipophilicity of a compound. The lipophilic property is according to the biodistribution of drugs. In the drug discovery, log P is also an important filter characteristic to screen the candidate of medications. The experimental methods usually need more chemicals, and cost time. Building a model can fast and massive estimate the log P values. One approach of data fusion was applied to fuse the predictive packages by traditional schemes and machine learning schemes in this study. In addition, the data manipulation retrieve also adopted to increase performance of traditional schemes. In traditional schemes, arithmetic mean (SUM) based on consistent data had the best statistical assessments, the statistical assessments gave r2 (0.91), rm2 (0.90), s (0.51), RMSE (0.51), ΔMax (3.63), and MAE (0.36); in machine learning schemes, deep neural network (DNN) showed the best results, the statistical parameters presented r2 (0.98), rm2 (0.97), s (0.26), RMSE (0.27), ΔMax (1.88), and MAE (0.16). Overall, traditional schemes can fastly compute the data to enhance the data screening, and machine learning schemes provide the better log P prediction.

1. Introduction 1
2. Materials and Methods 5
2.1 Data Sources 5
2.2 Predictive Packages 5
2.3 Data Fusion by Traditional Schemes 6
2.3.1 Traditional Schemes Based on All Data 6
2.3.2 Traditional Schemes Based on Consistent Data 6
2.3.3 Traditional Schemes Based on Purged Data 6
2.4 Data Fusion by Support Vector Regression 6
2.5 Data Fusion by Extreme Learning Machine 7
2.6 Data Fusion by Deep Neural Network 7
2.7 Data Partition 8
2.8 Predictive Evaluation 8
3. Results 11
3.1 Data Partition 11
3.2 Traditional Schemes 11
3.3 Machine Learning Schemes 13
4. Discussion 15
5. Conclusion 17
6. References 19

Andrés, A., Rosés, M., Ràfols, C., Bosch, E., Espinosa, S., Segarra, V., & Huerta, J. M. (2015). Setup and validation of shake-flask procedures for the determination of partition coefficients (logD) from low drug amounts. Eur. J. Pharm. Sci., 76, 181-191. doi:10.1016/j.ejps.2015.05.008
Arnott, J. A., Kumar, R., & Planey, S. L. (2013). Lipophilicity Indices for Drug Development J. Appl. Biopharm. Pharmacokinet., 1, 31-36 doi:10.14205/2309-4435.2013.01.01.6
Atkinson, H. C., & Begg, E. J. (1988). Relationship between Human Milk Lipid-Ultrafiltrate and Octanol-Water Partition Coefficients. J. Pharm. Sci., 77(9), 796-798. doi:10.1002/jps.2600770916
Austin, R. P., Davis, A. M., & Manners, C. N. (1995). Partitioning of ionizing molecules between aqueous buffers and phospholipid vesicles. J. Pharm. Sci., 84(10), 1180-1183. doi:10.1002/jps.2600841008
Avery, M. A., Bonk, J. D., Chong, W. K. M., Mehrotra, S., Miller, R., Milhous, W., . . . Wyandt, C. (1995). Structure-Activity Relationships of the Antimalarial Agent Artemisinin. 2. Effect of Heteroatom Substitution at O-11: Synthesis and Bioassay of N-Alkyl-11-aza-9-desmethylartemisinins. J. Med. Chem., 38(26), 5038-5044. doi:10.1021/jm00026a011
Bajusz, D., Rácz, A., & Héberger, K. (2019). Comparison of Data Fusion Methods as Consensus Scores for Ensemble Docking. Molecules, 24(15), 2690. doi:10.3390/molecules24152690
Belkin, N. J., Kantor, P., Fox, E. A., & Shaw, J. A. (1995). Combining the evidence of multiple query representations for information retrieval. Inf. Process. Manage., 31(3), 431-448. doi:10.1016/0306-4573(94)00057-A
Cheng, T., Zhao, Y., Li, X., Lin, F., Xu, Y., Zhang, X., . . . Lai, L. (2007). Computation of Octanol−Water Partition Coefficients by Guiding an Additive Model with Knowledge. Journal of Chemical Information and Modeling, 47(6), 2140-2148. doi:10.1021/ci700257y
Chi, C.-T., Lee, M.-H., Weng, C.-F., & Leong, M. K. (2019). In Silico Prediction of PAMPA Effective Permeability Using a Two-QSAR Approach. Int. J. Mol. Sci., 20(13), 3170. doi:10.3390/ijms20133170
Chirico, N., & Gramatica, P. (2012). Real External Predictivity of QSAR Models. Part 2. New Intercomparable Thresholds for Different Validation Criteria and the Need for Scatter Plot Inspection. J. Chem. Inf. Model., 52(8), 2044-2058. doi:10.1021/ci300084j
Clark, D. E., & Pickett, S. D. (2000). Computational methods for the prediction of ‘drug-likeness’. Drug Discovery Today, 5(2), 49-58. doi:10.1016/S1359-6446(99)01451-8
Cortes, C., & Vapnik, V. (1995). Support-vector networks. Mach. Learn., 20(3), 273-297. doi:10.1007/BF00994018
Dearden, J. C., & Bresnen, G. M. (1988). The Measurement of Partition Coefficients. Quant. Struct.-Act. Relat., 7(3), 133-144. doi:10.1002/qsar.19880070304
Ding, Y.-L., Lyu, Y.-C., & Leong, M. K. (2017). In silico prediction of the mutagenicity of nitroaromatic compounds using a novel two-QSAR approach. Toxicol. In Vitro, 40, 102-114. doi:10.1016/j.tiv.2016.12.013
Ghose, A. K., & Crippen, G. M. (1986). Atomic Physicochemical Parameters for Three-Dimensional Structure-Directed Quantitative Structure-Activity Relationships I. Partition Coefficients as a Measure of Hydrophobicity. J. Comput. Chem., 7(4), 565. doi:10.1002/jcc.540070419
Ginn, C. M. R., Turner, D. B., Willett, P., Ferguson, A. M., & Heritage, T. W. (1997). Similarity Searching in Files of Three-Dimensional Chemical Structures:  Evaluation of the EVA Descriptor and Combination of Rankings Using Data Fusion. J. Chem. Inf. Comput. Sci., 37(1), 23-37. doi:10.1021/ci960466u
Golbraikh, A., Shen, M., Xiao, Z., Xiao, Y.-D., Lee, K.-H., & Tropsha, A. J. J. o. C.-A. M. D. (2003). Rational selection of training and test sets for the development of validated QSAR models. J. Comput.-Aided Mol. Des., 17(2), 241-253. doi:10.1023/a:1025386326946
Hall, D. L., & Llinas, J. (1997). An introduction to multisensor data fusion. Proc. IEEE, 85(1), 6-23. doi:10.1109/5.554205
Hansch, C., Leo, A., & Hoekman, D. (1995). Exploring QSAR: Hydrophobic, Electronic, and Steric Constants (Vol. 2). Washington, DC: American Chemical Society.
Harnisch, M., Möckel, H. J., & Schulze, G. (1983). Relationship between log Pow, shake-flask values and capacity factors derived from reversed-phase high-performance liquid chromatography for n-alkylbenzenes and some oecd reference substances. J. Chromatogr. A, 282, 315-332. doi:10.1016/S0021-9673(00)91610-8
Huang, G.-B., Zhu, Q.-Y., & Siew, C.-K. (2004, 25-29 July 2004). Extreme learning machine: a new learning scheme of feedforward neural networks. Paper presented at the 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541).
Huang, G.-B., Zhu, Q.-Y., & Siew, C.-K. (2006). Extreme learning machine: Theory and applications. Neurocomputing, 70(1), 489-501. doi:10.1016/j.neucom.2005.12.126
Huang, G., Zhou, H., Ding, X., & Zhang, R. (2012). Extreme Learning Machine for Regression and Multiclass Classification. IEEE Trans. Syst. Man Cybern. Part B-Cybern., 42(2), 513-529. doi:10.1109/TSMCB.2011.2168604
Hubert, M., & Engelen, S. (2004). Robust PCA and classification in biosciences. Bioinformatics, 20(11), 1728-1736. doi:10.1093/bioinformatics/bth158 %J Bioinformatics
Kearsley, S. K., Sallamack, S., Fluder, E. M., Andose, J. D., Mosley, R. T., & Sheridan, R. P. (1996). Chemical Similarity Using Physiochemical Property Descriptors. J. Chem. Inf. Comput. Sci., 36(1), 118-127. doi:10.1021/ci950274j
Kril, M. B., & Fung, H.-L. (1990). Influence of hydrophobicity on the ion exchange selectivity coefficients for aromatic amines. J. Pharm. Sci., 79(5), 440-443. doi:10.1002/jps.2600790517
Lee, M.-H., Ta, G. H., Weng, C.-F., & Leong, M. K. (2020). In Silico Prediction of Intestinal Permeability by Hierarchical Support Vector Regression. Int. J. Mol. Sci., 21(10), 3582. doi:10.3390/ijms21103582
Leeson, P. D., & Springthorpe, B. (2007). The influence of drug-like concepts on decision-making in medicinal chemistry. Nat. Rev. Drug Discov., 6(11), 881-890. doi:10.1038/nrd2445
Leo, A., Hansch, C., & Elkins, D. (1971). Partition coefficients and their uses. Chem. Rev., 71(6), 525-616. doi:10.1021/cr60274a001
Leong, M. K., Chen, Y.-M., & Chen, T.-H. (2009). Prediction of human cytochrome P450 2B6-substrate interactions using hierarchical support vector regression approach. J. Comput. Chem., 30(12), 1899-1909. doi:10.1002/jcc.21190
Leong, M. K., Lin, S.-W., Chen, H.-B., & Tsai, F.-Y. (2010). Predicting Mutagenicity of Aromatic Amines by Various Machine Learning Approaches. Toxicol. Sci., 116(2), 498-513. doi:10.1093/toxsci/kfq159 %J Toxicological Sciences
Liao, Q., Yao, J., & Yuan, S. (2006). SVM approach for predicting LogP. Mol. Diversity, 10(3), 301-309. doi:10.1007/s11030-006-9036-2
Lipinski, C. A., Lombardo, F., Dominy, B. W., & Feeney, P. J. (1997). Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews, 23(1), 3-25. doi:10.1016/S0169-409X(96)00423-1
Meylan, W. M., & Howard, P. H. (2000). Estimating log P with atom/fragments and water solubility with log P. Perspect. Drug Discovery Des., 19(1), 67-84. doi:10.1023/A:1008715521862
Molnár, L., Keserű, G. M., Papp, Á., Gulyás, Z., & Darvas, F. (2004). A neural network based prediction of octanol–water partition coefficients using atomic5 fragmental descriptors. Bioorg. Med. Chem. Lett., 14(4), 851-853. doi:10.1016/j.bmcl.2003.12.024
Moriguchi, I., Hirono, S., Liu, Q., Nakagome, I., & Matsushita, Y. (1992). Simple Method of Calculating Octanol/Water Partition Coefficient. CHEM. PHARM. BULL., 40(1), 127-130. doi:10.1248/cpb.40.127
Ojha, P. K., Mitra, I., Das, R. N., & Roy, K. (2011). Further exploring rm2 metrics for validation of QSPR models. Chemometrics Intell. Lab. Syst., 107(1), 194-205. doi:10.1016/j.chemolab.2011.03.011
Ranadive, S. A., Chen, A. X., & Serajuddin, A. T. M. J. P. R. (1992). Relative Lipophilicities and Structural-Pharmacological Considerations of Various Angiotensin-Converting Enzyme (ACE) Inhibitors. Pharm. Res., 9(11), 1480-1486. doi:10.1023/a:1015823315983
Roy, K., Mitra, I., Kar, S., Ojha, P. K., Das, R. N., & Kabir, H. (2012). Comparative Studies on Some Metrics for External Validation of QSPR Models. J. Chem Inf. Model., 52(2), 396-408. doi:10.1021/ci200520g
Sangster, J. (1989). Octanol‐Water Partition Coefficients of Simple Organic Compounds. J. Phys. Chem. Ref. Data, 18(3), 1111-1229. doi:10.1063/1.555833
Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural Netw., 61, 85-117. doi:10.1016/j.neunet.2014.09.003
Sheridan, R. P., Miller, M. D., Underwood, D. J., & Kearsley, S. K. (1996). Chemical Similarity Using Geometric Atom Pair Descriptors. J. Chem. Inf. Comput. Sci., 36(1), 128-136. doi:10.1021/ci950275b
Teijeiro, S. A., Moroni, G. N., Motura, M. I., & Briñón, M. C. (2000). Lipophilic character of pyrimidinic nucleoside derivatives: correlation between shake flask, chromatographic (RP-TLC and RP-HPLC) and theoretical methods. J. Liq. Chromatogr. Relat. Technol., 23(6), 855-872. doi:10.1081/JLC-100101494
Tropsha, A., Gramatica, P., & Gombar, V. K. (2003). The Importance of Being Earnest: Validation is the Absolute Essential for Successful Application and Interpretation of QSPR Models. QSAR Comb. Sci., 22(1), 69-77. doi:10.1002/qsar.200390007
Valkó, K. (2004). Application of high-performance liquid chromatography based measurements of lipophilicity to model biological distribution. Journal of Chromatography A, 1037(1), 299-310. doi:10.1016/j.chroma.2003.10.084
Willett, P. (2006). Enhancing the Effectiveness of Ligand-Based Virtual Screening Using Data Fusion. QSAR Comb. Sci., 25(12), 1143-1152. doi:10.1002/qsar.200610084
Willett, P. (2013). Combination of Similarity Rankings Using Data Fusion. J. Chem. Inf. Model., 53(1), 1-10. doi:10.1021/ci300547g